As Agile practices and the DevOps movement transform delivery pipelines throughout the world, let’s take a moment to reflect on the database administrator’s role in pushing the latest features out the door. Here are six best practices for users of XL Deploy:
1. Remember that the DBA is part of the application development team(s).
The DBA coaches the developers on database architecture, proper SQL standards, Explains, etc. as they design, build, and test their code. All team members operate under Agile principles.
2. Standard environments that include the database should be able to be stood up quickly for testing the latest build.
In XL Deploy: Include a dedicated database in every Environment to which the application is deployed
3. Database configuration (DDL, SQL) is developed and saved as version-controlled code.
To paraphrase Gene Kim, Infrastructure-as-Code allows modern development practices to applied to the entire development stream to enable fast deployment, continuous integration, continuous delivery, and continuous deployment.
4. Database changes are aligned with the application changes that need them.
In XL Deploy, an Application Package is a complete and environment-independent package, able to be deployed to any environment. By complete, we mean that It includes all the binary artifacts, configuration settings, startup commands, etc. to get the application properly configured and running. A dictionary/placeholder scheme is used to substitute environment-dependent values (port numbers, credentials, etc.) at deployment time.
For example, an application needs a new index on a table, so the DDL for the index is part of the application change package to ensure that both are always deployed together.
In XL Deploy: Include sql.SqlScripts element in the application package. This element is simply a zip file of lexicographically ordered scripts. See below for an example.
Database configurations used by more than one application can be managed by using XL Deploy’s application dependencies set at the udm.DeploymentPackage object. See https://docs.xebialabs.com/xl-deploy/concept/application-dependencies-in-xl-deploy.html.
5. Database configuration changes are accumulated in new versions of the sql.SqlScripts object as they are developed.
At any point in time, the object can be deployed to a new or old environment and bring the database up to the state necessary for that version of the application. So, in a brand new environment with just the DBMS software, XL Deploy would run all the scripts. In an environment with a previous deployment(s), XL Deploy runs only scripts not run previously.
Here is a simple example of the evolution of the scripts in sql.Scripts.Version 1 — this version configures a database from the bare metal:
Version 4: — this iteration drop the bad index. It may seem redundant to create and drop the index in the same set of scripts, but remember that this package may update an Environment whose current state is at Version 2 and we want to make we drop the bad index.
6. Rollback scripts are included to allow backout and undeployment.
XL Deploy uses a regex pattern to recognize which rollback script corresponds to which forward-going script. Developers and DBAs should include these as they build the package of change. Each one undoes the forward-going action, e.g. the first one drops the tables created by the first script. With the rollbacks included, the package should look like this:
Through NBI’s approach to be the best open-access operator in telecom, they have been supported by key technology partners, with automation capabilities provided by Digital.ai Release and Deploy underpinning integral aspects.
Regardless of where you are on your digital transformation journey, we can help you achieve your strategic outcomes and accelerate value delivery with the right combination of technology, services, and training.